Chancellor, R. D. & B.-U. Meyburg eds. 2004 Raptors Worldwide WWGBP/MME

Phylogenetic Relationships in the Hierofalco Complex (Saker-, Gyr-,

Lanner-, Laggar Falcon)

Michael Wink, Hedi Sauer-Giirth, David Ellis & Robert Kenward

ABSTRACT Saker, Gyr, Lanner and Laggar represent large falcons that belong to the

Hierofalco complex. We have amplified and sequenced the mitochondrial cytochrome b gene of all representatives of this clade, including several hundreds of Sakers from Kazakhstan and Mongolia. DNA data show that these taxa are very closely related. Within the Saker falcons 3 genetic lineages can be detected. No evidence could be obtained that the Altai falcon or the other Saker subspecies can be recognised at the DNA level. Within the Lanner falcon two lineages are evident that correspond to the subspecies F. b. biarmicus or F. b. feldeggii. The Australian F. subniger appears to be associated with the Hierofalco complex.

INTRODUCTION Although Gyr, Lanner and Saker falcons are apparently closely related, they

show a high degree of plumage polymorphism and have puzzled taxonomists for some time. 0. Kleinschmidt (1901, 1958) and Meinertzhagen (1954) had placed them in a single species, Falco hierofalco, and distinguished F. h. islandus Brünn, F. h. norvegicus Schleg., F. h. cherrug Gray and F. h. feldeggi Schleg. F. hierofalco has been split since and today some authorities include F. rusticolus, F. cherrug, F. biarmicus, F. jugger and F. mexicanus in the subgenus Hierofalco and consider them as a superspecies (Amadon & Bull 1988; del Hoyo et al. 1994; Sibley & Monroe 1990; Cramp 1980). Within the species some subdivision, mostly concerning geographical forms or even morphs, has been considered (Niethammer 1938; Dementiev 1960; Weick 1980; Amadon & Bull 1988; del Hoyo et al. 1994; Sibley & Monroe 1990):

499

• F. rusticolus L.: Several subspecies; F. r. rusticolus (Scandinavia and northern Russia); F. r. uralensis Menzbier (Russia east of Petschora); F. r. islandus Brünnich (on Iceland); F. r. candicans Gmelin (on Greenland)

• F. cherrug: Several subspecies; F. c. cherrug Gray (E Europe, Russia to Altai mountains and Yenisey) and F. c. milvipes Jerdon (Siberia, Mongolia, China) (del Hoyo et al., 1994). Weick (1980) distinguishes: F. c. cherrug, F. c. cyanopus Thiemann (Bohemia to Wolga River), F. c, saceroides (Bianchy) (Kirghiz steppe Saja, Altai, Tarbatagai, NW Mongolia), F. c. coatsi Dementiev (Tian Shan west to NE Iran), F. c. hendersoni Hume (Pamir, Himalayas to Kansu), F. c. milvipes Jerdon (Mongolia, Manchuria). Weick includes F. c. altaicus in F. c. milvipes and also concludes that F. c. saceroides, F. c. coatsi and F. c. hendersoni could be colour morphs and would then be treated as F. c. milvipes.

• F. altaicus Menzbier: a dark coloured falcon of central Asia (Altai region, NW Mongolia, W China); treated also as a morph, either closer to F. rusticolus or F. cherrug. Ellis (1995) provided arguments that F. altaicus is a good species.

• F. biarmicus: 5 geographically defined subspecies; F. b. biarmicus Temminck (southern Africa), F .b. abyssinicus Neumann (western, central and eastern Africa), F. b. tanypterus (northern Africa, Israel, Iraq), F. b. erlangen Kleinschmidt (NW Africa), F. b. feldeggii Schleg. (southern Italy, E to Armenia and Azerbaijan)

• F. jugger : monotypic (Indian subcontinent) • F. mexicanus: (North America) We have started to use molecular genetic techniques to elucidate the

complex relationships in the Hierofalco complex. We mainly rely on nucleotide sequences of the mitochondrial cytochrome b gene that shows a good resolution on genus and even species level. Nucleotide sequences of the mitochondrial cytochrome b gene have already been employed to study the systematics and evolution of diurnal raptors (Griffiths 1997; Mindell 1997; from our laboratory: Groombridge et al. 2002, Kruckenhauser et al. 2003; Riesing et al. 2003; Seibold et al. 1993, 1996; Seibold & Helbig 1995, 1996; Helbig et al. 1994; Wink 1995, 1998, 2000; Wink & Seibold 1996; Wink & Sauer-Gürth 2000; Wink et al. 1996, 1998a,b;).

First and preliminary results have been published already on the phylogenetic relationships of Hierofalcons. The first publications based on the PhD thesis of I. Seibold (Seiboldl994) (Seibold et al. 1993; Wink 1995b, 1998; Wink & Seiboldl996; Wink et al. 1998) presented a mistaken picture, since nuclear copies of the cytochrome b genes had been sequenced by chance (in Hierfalcons but not for the other falcons), which placed the Hierofalcons at the base of the falcon tree. The mistake was identified when new PCR primers and new sequencing techniques were employed. Later phylogenies unequivocally showed that F. rusticolus, F. biarmicus, F. cherrug and F. jugger form a closely related complex that clusters as a sister group to the peregrines (Wink 2000; Wink & Sauer-Gürth 2000). F. mexicanus is not a member of this clade but shares ancestry with both peregrines and hierofalcons. A closer relationship with peregrines had been postulated by Schmutz & Oliphant (1987).

500

We have amplified and sequenced the mitochondrial cytochrome b gene of all representatives of this clade, including several hundreds of Sakers from Kazakhstan and Mongolia. Samples are included that were identified by scientists and falconers as "Altai falcons". Although the present study is still incomplete, since important samples from several populations/subspecies are still missing, a few conclusions can be drawn and a new hypothesis put forward on the phylogeny and taxonomy of this falcon complex.

MATERIAL AND METHODS We have isolated total DNA from feather, blood or tissue samples (see

Wink 2000) which had been kindly supplied by several colleagues (W. Bednarek, J. Bragin, F. Nittinger, A. Gamauf, R. Pfeffer, D. Ristow, J. Penhallurick, M. Heidenreich). The cytochrome b gene was amplified by PCR (primer sequences in Wink & Sauer-Gürth 2000) and sequenced by using AlfExpress (Amersham Pharmacia Biotech) or a capillary sequencer ABI 3100 (Applied Biosystems) instruments. Sequences of 1000 and more base pairs were aligned manually and analysed with the software packages PAUP* (Swofford, 2002) and MEGA2 (Kumar et al. 2001) (see Wink 2000; Wink & Sauer-Gürth 2000; Wink et al. 2002; Broders et al. 2003 for further details).

RESULTS AND DISCUSSION Figure 1 shows a molecular phylogeny of the Hierofalco complex based on

a reconstruction with Maximum Parsimony (NJ and ML lead to similar results; trees are not shown because of limited space). The overall topology shows that:

• The peregrines (F. peregrinus, F. pelegrinoides) form a sister group to the Hierofalcons that include F. rusticolus, F. cherrug, F. biarmicus, F. jugger and apparently F. subniger.

• F. mexicanus is not a member of the Hierofalco group sensu stricto; thus Schmutz & Oliphant (1987) were right in their conclusions that F. mexicanus is closer to F. peregrinus than to F. rusticolus.

• F. subniger from Australia clusters at the base of the Hierofalco clade; sequences of more birds are needed to corroborate this surprising finding. Since members of the Hierofalco complex were found on all continents except Australia, this finding would make sense.

• Members of the Hierofalco clade are closely related; the interspecific genetic distances between Gyr, Saker and Lanner falcons are in the range of 0.4 to 2.0%, indicating that they have evolved during the last 200,000 to 1 million years (assuming a 2% divergence = 1 million years; Wilson et al. 1987; Tarr & Fleischer 1993). This is the divergence range of young species; in some bird families taxa with such small genetic distances would be considered as subspecies. Thus the concept of O. Kleinschmidt (1901) or Meinertzhagen (1954) to lump these taxa into a single species Falco hierofalco had some merits and justification.

Our data allow some comments on the intraspecific genetic variation of F. rusticolus, F. cherrug and F. biarmicus:

• The Gyr falcons appear as a monophyletic group, often clustering at the base of the Gyr/Saker/Lanner assemblage but no birds from Siberia or North America have been studied so far.

• Lanner falcons are represented in our study by two subspecies: F. b.

501